Updated backup cuts

[apcooper]

This PR updates the backup target selection cuts to match the backup program document: https://www.overleaf.com/project/61116134e46f88886a3367d2

• Add bp-rp based mag thresholds
• Add new backup_giant target class

[coveralls]

Coverage increased (+0.05%) to 57.835% when pulling 4df593cf1b2d32e186d17dd4bbd555ffa01d2200 on backup into bb5d5c99f187631c3ac9d4a1eb60716a6076d164 on master.

[apcooper]

This seems to work as expected in selecting backup targets. Around the pole the density is ~900 per sq. deg. Targets are selected ok outside the footprint, and the zone of avoidance around the plane looks ok for BACKUP_FAINT and BACKUP_VERY_FAINT.

[apcooper]

Quick Gaia CMD of the selection around the NGP, green/red/blue are bright/faint/very faint, 'magma' is giants. I don't think the artificial stripe at BP-RP = 0 is a problem.

[apcooper]

@geordie666 I think this is ready for you to take a look, thanks.

[apcooper]

Many thanks for the review and comments @geordie666. Those sound like very good reasons to keep the in_gal cut. The MWS already signed off on that, so we're fine as we are. I fixed changes.rst.

Please let me know when you generate targets with these selections so I can ask @araichoor can run fiber assignment on them.

[geordie666]

I'm going to leave this open a little longer; @schlafly wants to check some bright-end cuts in G, Bp, Rp, in case of saturation. So we may need to add a few more lines of code before running this.

[schlafly]

This looks good to me. Here's the grz magnitudes according to Rongpu's relations at the bright cut defined by the backup program, G > 11.2 + 0.6*(BP-RP):

If we trust this then we'd like to be a bit more careful with the very blue sources with BP-RP ~ -1. These are vanishingly rare. I'd propose three additional cuts: G > 11.2 BP > 11.2 RP > 11.2 These touch very few objects that aren't already cut out by the existing sliding cut, while also giving us some security in case an object has truly bizarre colors.

With this addition the brightest g mags from Rongpu passing the cut are at ~11th mag at BP-RP ~ -0.5.

[geordie666]

@apcooper : Do Eddie's suggested bright-end cuts look good? If so, I'll add those cuts, merge this PR, and then run an example set of targets. Cori is down tomorrow, so it might be Thursday before we have the files.

[apcooper]

Thanks. Looks OK but I'd like @segasai's opinion.

[segasai]

The reason why I think the original cut is sufficient is because the G<A+B*(BP-RP) was determined based on the actual count-rate vs BP-RP. And for those extremely blue objects the count rate is not approaching anything close to saturation. Here is the plot show the limiting magnitude as a function of BP-RP for different chips. And black line is the 11.2+.6.(Bp-RP).

However despite that I agree that here we may want to be safer. So I think, those proposed are fine

[geordie666]

Thanks, all. Unless I hear otherwise, once Cori is back up (or more likely tomorrow) I'll add Eddie's proposed bright limits and merge this PR.

@apcooper: As Cori will be back up sooner, your-time (I think), feel free to beat me to this and add Eddie's limits yourself!

[apcooper]

OK, thanks, I made that change. @geordie666, please merge when you're ready.

[geordie666]

Thanks @apcooper. I updated the wiki to include the new bright-end cuts. I'll merge once tests pass and run a full suite of example targets.